I'm surprised that you even leap to people like Hans Zimmer and others.
The people we need to worry about are aallll of the people earning a living for everything else like background music for Indi games, Ambient music etc.
"Godwin himself has also criticized the overapplication of the law, claiming that it does not articulate a fallacy, but rather is intended to reduce the frequency of inappropriate and hyperbolic comparisons. "
I don't think the original comparison was inapt -- the video is hard evidence of appalling and disgusting propaganda. It will not be viewed kindly by history.
A comparison doesn’t mean a literal equivalence. It means a similarity. In this case there is a similarity with regards to the propaganda. They’re not equivalent but there are definitely not unlike each other either.
I think the gp is kind to even call this "hyperbolic". There is no comparison between Amazon and a concentration camp, just because people are worked hard at amazon.
The reason Auschwitz was bad was not because people worked hard there. It was because people were enslaved and murdered there.
It would be like comparing a camp site to a concentration camp, because both are "camps". The comparison yields no useful insights.
But the point of the comparison is not to make Amazon/Bezos out to be Nazis; it's to point out that this practice might in the future be looked upon as an artefact of the times that hopefully no longer happen in the museum visitor's time.
In general, fine-tuning helps a model figure out how to do the exact task that is being done in the examples it's given. So fine-tuning it on 1000 examples of an API being used in the wild is likely to teach it to use that API really effectively, but fine-tuning it on just the API docs probably won't.
That said, there are a lot of interesting ideas floating around on how to most effectively teach a model purely from instructions like API docs. Powerful models like GPT-4 can figure it out from in-context learning (ie. if you paste in a page of API docs and ask GPT-4 to write something with the API it can usually do a decent job). I suspect the community will figure out techniques either through new training objectives or synthetic training data to do it for smaller fine-tuned models as well.
Generally speaking, fine-tuning a small model makes sense when the task that you want it to carry out is well-defined and doesn't vary too much from one prompt to another. Fine-tuning facts into a model doesn't seem to scale super well, but general textual style, output format, and evaluation criteria for example can all be instilled through the fine-tuning process. I would use lookup if you need your answers to include a wide array of information that the model you're basing off of wasn't initially trained on.
Nonetheless for just getting to your goal and not becoming really good in something this is not necessary or can be done later